oneflow.cuda

is_available	Returns a bool indicating if CUDA is currently available.
device_count	Returns the number of GPUs available.
current_device	Returns local rank as device index.
set_device	Sets the current device.
synchronize	Waits for all kernels in all streams on a CUDA device to complete.

Note

The current_device returns local rank as device index. It is different from the ‘torch.current_device()’ in PyTorch.

Random Number Generator

manual_seed_all	Sets the seed for generating random numbers on all GPUs.
manual_seed	Sets the seed for generating random numbers for the current GPU.

GPU tensor

HalfTensor	The tensortype oneflow.cuda.HalfTensor is not available.
FloatTensor	The tensortype oneflow.cuda.FloatTensor is not available.
DoubleTensor	The tensortype oneflow.cuda.DoubleTensor is not available.
BoolTensor	The tensortype oneflow.cuda.BoolTensor is not available.
ByteTensor	The tensortype oneflow.cuda.ByteTensor is not available.
CharTensor	The tensortype oneflow.cuda.CharTensor is not available.
IntTensor	The tensortype oneflow.cuda.IntTensor is not available.
LongTensor	The tensortype oneflow.cuda.LongTensor is not available.